Here’s a high-class question for you: “If your brand was going out for the evening, to a live performance with its friends, who would they be going to see together?”
And here’s a low-grade one: “Is it time to refresh the background image on the campaign banners?”
Traditionally, data in marketing has been an aid to decision making. Now, increasingly, we’re applying machine learning. It does rely on the same data we’ve always used - a lot of it - but it’s a new capability, that works on another level. Machine learning is for eliminating the need to make decisions at all.
In strategic, creative and display media teams in digital marketing, we‘ll see the type of work we do evolve. If we capitalise on machine learning technology, we’ll be able to pay less attention to designing and producing creative variants, and to planning and analysing tests; and we’ll spend more time and energy on the higher-class questions, like who the brand is seeking to engage, and how those people might like to be spoken to.
Smarter speech recognition
Speaking of speech, it’s the other main area where we’ll see the impact of machine learning in creative. We’ve seen recent dramatic improvements in the technology that understands and synthesises natural language and speech. These have been made possible by machine learning. Put simply, the big deal with this new tech - for you and me - in strategy, creative and media teams - is that soon it will be common for a digital ad to hold actual individual conversations with each of the people it reaches.
Digital advertising has of course long attempted to be personal… or personalised, at least. As we understand intuitively, if an ad is relevant for a person, it will be able to communicate with them, and persuade. Being relevant - more relevant than everything else that’s trying to speak to that person - is also a good way to win lasting attention. Attention makes ads more competitive and media spends more efficient.
So, being relevant is the key: to effectiveness, competitiveness and efficiency. The problem is, it’s a shot in the dark. I mean, honestly, if we expect to deliver a specific sales message with any sort of volume, how can we expect it to be relevant? We’ll be lucky, really, if we know exactly what’s going on with any one of these people’s lives right now. They’re almost all probably trying just to get on with something else. An advert isn’t personal - for goodness’ sake - it doesn’t even… talk.
As personalisation meets interaction, I think we might just at last have reached comms nirvana. We’ll still have planning to do - planning what to offer, to whom and how - but instead of having also to plan and manage tests and account structures, we’ll apply our strategy directly, by embodying it (virtually, for now) in creative that can speak for us.
The capacity for dialogue will sometimes be worth more than all the ad variant combinations you could possibly generate. The irony of traditional ad creative is that the question of whether or not it’s effectively personalised isn’t actually relevant. Traditional ads can’t progress past the point of a handshake, because they can’t listen and interact. For humans, when we meet someone, whether there’s been a personal connection or not - whether we understand one another - is something we judge (if we’re being rational) on the basis of how we’ve interacted.
And, returning to the idea of eliminated decisions: as machine learning impacts on creative in digital marketing, you won’t be losing any sleep over keeping your conversation-capable creative optimised. It’ll vary and optimise itself. Instead, you can lose sleep over much bigger things, like whether your brand is ready to listen to what the world has to say.
Kevin Joyner is director of planning and insight at Croud